Last weekend wasn’t a total waste, although the problems at Blue Host were somewhat trying … meanwhile, have implemented updates to the way the RetroRead “colophon” elements are injected into the ebook. Previously I was using some assumptions about the canonical layout of a Google-generated ebook which, while correct in what still seems like the majority of cases, don’t hold up universally, with the result that in some cases, I was over-writing some of the introductory material. I now do a “smart” injection of the RetroRead title page and colophon text such that it is directly context-sensing, rather than making assumptions. Result: better books!
Meanwhile, need to look at either batch re-processing existing titles to re-fresh the conversions, or (more likely), tweak the “file exists” logic to allow the original uploader of a title to re-do it and replace (I already have this implemented for my own account for test purposes.)
